adambeynon changed the topic of #opal to: - Ruby runtime and library on top of Javascript | 1.0.0 is near | This channel is logged at
<wmnnd_> I have encountered a little incompatibility between standard Ruby and Opal. Any suggestions on how to deal with it? =>
<meh`> wat
<meh`> wmnnd_, master or latest stable?
<meh`> because it ain't happening here
<meh`> (on master)
<wmnnd_> Latest stable.
<meh`> yeah, there have been A LOT of commits between latest stable and master
<meh`> the release is next door
<meh`> wmnnd_, you might want to run off master
<meh`> until 0.6 is released
<wmnnd_> Can I tell bundler to do that?
<meh`> yes
<meh`> just gem 'opal', github: 'opal/opal'
<meh`> if you're using native, you have to require 'native'
<meh`> that's basically the only backward incompatible change
<meh`> IIRC
<meh`> but you were already doing that reading the pastebin
<wmnnd_> Thank you, meh`. I think I'll also give Vienna a shot then. But it's late now, so I'll do that tomorrow.
<wmnnd_> Thank you for your help so far, but I shall return :-)
wmnnd_ has quit [Quit: leaving]
harrow has quit [Ping timeout: 264 seconds]
harrow has joined #opal
lacrosse has quit [Ping timeout: 245 seconds]
lacrosse has joined #opal
adambeynon_ has quit [Ping timeout: 272 seconds]
adambeynon_ has joined #opal
GitHub88 has joined #opal
<GitHub88> opal/master 3619bd8 meh: Remove String#to_a since it shouldn't exist
GitHub88 has left #opal [#opal]
<GitHub88> [opal] meh pushed 1 new commit to master:
travis-ci has joined #opal
<travis-ci> [travis-ci] opal/opal#1676 (master - 3619bd8 : meh): The build passed.
<travis-ci> [travis-ci] Build details :
travis-ci has left #opal [#opal]
GitHub5 has joined #opal
GitHub5 has left #opal [#opal]
<GitHub5> [opal] meh pushed 2 new commits to master:
<GitHub5> opal/master 3134587 meh: Add Numeric#lcm and Numeric#gcd
<GitHub5> opal/master bf01576 meh: Alias Numeric#kind_of? or it won't use the current one
GitHub130 has joined #opal
<GitHub130> [opal] meh created numeric (+1 new commit):
GitHub130 has left #opal [#opal]
<GitHub130> opal/numeric ac2c7c6 meh: WIP
meh` has quit [Quit: I don't want to live on this planet anymore.]
travis-ci has joined #opal
<travis-ci> [travis-ci] opal/opal#1677 (master - bf01576 : meh): The build passed.
<travis-ci> [travis-ci] Build details :
travis-ci has left #opal [#opal]
travis-ci has joined #opal
<travis-ci> [travis-ci] Build details :
<travis-ci> [travis-ci] opal/opal#1678 (numeric - ac2c7c6 : meh): The build passed.
travis-ci has left #opal [#opal]
<travis-ci> [travis-ci] Change view :
ryanstout has quit [Quit: ryanstout]
marcandre has quit [Remote host closed the connection]
GitHub67 has joined #opal
<GitHub67> [opal-jquery] adambeynon pushed 1 new commit to master:
GitHub67 has left #opal [#opal]
<GitHub67> opal-jquery/master 1a4a76b Adam Beynon: Add shorted event control names
GitHub131 has joined #opal
<GitHub131> [opal-jquery] adambeynon pushed 1 new commit to master:
GitHub131 has left #opal [#opal]
<GitHub131> opal-jquery/master 3c7dca1 Adam Beynon: whitespace
kludge` has quit [Ping timeout: 245 seconds]
kludge` has joined #opal
ryanstewart has quit [*.net *.split]
yazgoo has quit [*.net *.split]
adambeynon_ has quit [*.net *.split]
adambeynon_ has joined #opal
ryanstewart has joined #opal
yazgoo has joined #opal
kings has joined #opal
yazgoo has quit [Ping timeout: 243 seconds]
yazgoo has joined #opal
yazgoo has quit [Ping timeout: 248 seconds]
yazgoo has joined #opal
<kings> morning
<kings> adambeynon_: I'm looking at opal-rspec - in the example I don't see any of the rspec api? the should etc
<adambeynon_> kings: I think should() is deprecated in favor of the expect(foo).to() syntax
<adambeynon_> kings: should still works though
yazgoo has quit [Ping timeout: 272 seconds]
<kings> adambeynon_: ah I just read a blog post about expect - I see - I better get with the times!!
<adambeynon_> kings: lol. I still prefer the should() syntax as well, and still use it with all my code
<kings> I've been using should for years - and also in Java and Scala - but ruby does tend to drive innovation
<kings> which is good :)
<kings> im also dragonkh btw - I forgot to log out on the other computer
<adambeynon_> ah right. how are you getting on with opal?
<adambeynon_> I have actually been writing some docs in last few days, should help out :)
<kings> adambeynon_: yeah its going well - It's opened up a whole new set of possibilities which is totally aweseome - I'm not that keen on js - so I have been using cofeescript but didnt like that much either so switching back into ruby is amazing
<kings> I had a look at opal a few years ago but for some reason didn't get excited about it
<kings> I think now it's probably because lots of web dev has gone further towards js than ever before
<kings> I've been a ruby fanatic for 12 years now - so if I have to write js - I'm certainly writing it with opal :)
<kings> I'm looking at vienna also - looks good - I had a look at lissio also
yazgoo has joined #opal
<adambeynon_> kings: I think its fair to say that opal has only really been "production ready" in the last 12 months or so
<adambeynon_> the generated code for starters is now a LOT better
<kings> adambeynon_: how long have you been working on opal - did you author it from the start?
<adambeynon_> kings: the main difference between vienna and lissio is that vienna uses opal-jquery, and lissio uses opal-browser
<adambeynon_> kings: yes. about 4 years or so
<kings> yeah I saw - I prefer opal-jquery
<kings> you have been very busy over the last few months then - I see lots of stuff - did you get some motivation! lol
<adambeynon_> kings: lol, well its only the last 2 years that I've actually managed to use it in some production apps
<adambeynon_> which digs up lots of bugs which need fixing :)
<kings> thats pretty cool
<kings> what kind of apps do you have it running in?
<adambeynon_> I have 6. 2 are just light sprinkles with opal-jquery, the other 4 are full client side mvc
<adambeynon_> using opal-haml for view rendeirng
<kings> nice!
<adambeynon_> rails based json api backend
<kings> they public?
<adambeynon_> unfortunately not. all for internal/private systems
<adambeynon_> kings: I keep meaning to try and write a larger example app
<adambeynon_> the todos example app is really basic
<adambeynon_> doesnt show off opal-jquery or vienna to their abilities
<kings> could be interesting for mobile apps - with zepto
<kings> well I better get coding - my wife wants us to go out - and my baby girl is scampering about the floor - who know when I'll get half an hour to code again lol !!!!
<kings> thanks for answering my questions :)
<adambeynon_> kings: np :)
<kings> adambeynon_: can I run opal-rspec without phantomjs?
<kings> without running in the browser?
<adambeynon_> kings: in node.js, for example?
<kings> no just in a plain opal app
<kings> when I try to run with phantomjs - it just hangs
<kings> its hard to see what is going on
<adambeynon_> kings: hangs for how long? does it finish?
<kings> not sure still waiting - 1 min so far
<kings> also the example doesnt run in opal-jquery - it's says it missing the opal gem
<kings> havent investigated the example much though - its probably just needs gem 'opal' or something in the gemfile
<kings> adambeynon_: says Specs timed out
meh` has joined #opal
<kings> I've just got a very simple spec: expect(1).to eq(1)
<kings> I think its something to do with the paths in my Rakefile though
<adambeynon_> kings: how are you running the specs?
<adambeynon_> with the Rake task?
<kings> I've tried 2 ways:
<kings> 1. with a rake task - which hangs then Times out
<kings> 2. with a rack app - which just says: 0 examples, 0 failures (time taken: 0.004)
<kings> in the browser
<kings> I have the spec folder just in the root of my project
<kings> and my application.rb is in an app dir at the root of the project
<kings> all my classes are in the app dir as well
<kings> this is my Rakefile:
<adambeynon_> kings: I have never experienced the spec runner just hanging
<adambeynon_> perhaps a phantomjs version issue?
<adambeynon_> what version phantomjs are you running?
kings_ has joined #opal
<kings_> sorry computer crahded
kings has quit [Ping timeout: 272 seconds]
<kings_> grr stupid computer
<meh`> adambeynon_, how do I hack the superclass for Number?
<adambeynon_> meh`: I dont get what the benefit of this is
<meh`> Math won't work otherwise
<meh`> or the benefit of having Numeric compliant?
<adambeynon_> Numeric being compliant
<meh`> I think you're looking at it from your specific perspective
<meh`> you don't use much of the Ruby features in your stuff
<meh`> being able to run Ruby gems unmodified is very important
<meh`> and we need compliancy for that
<meh`> and other than that, we need compliant Numeric if we want to do anything with numbers
<meh`> if your concern is "the size becomes bigger, and *I* don't even need it"
<meh`> it's not a good argument either
<meh`> if your concern is about size, you're working on mopal
<meh`> we should be compliant where possible
<kings_> adambeynon_: can't get my specs to run :(
<meh`> and hacking a superclass isn't that much of an issue
<adambeynon_> meh`: compliant where it makes sense - we are compiling to javascript which has a single class for numbers
<adambeynon_> rational + complex I agree with
<adambeynon_> those can/should be added
<meh`> adambeynon_, you're missing the point
<meh`> there won't be any change to Number
<meh`> at all
<meh`> it's only a hack to have Numeric as parent
<meh`> everything else depends on the idea that Numeric is the parent class for any kind of number
<meh`> Math depends on that fact, and other gems depend on those semantics as well
<meh`> anything that has to do with numbers or things that act as numbers will be broken without a Numeric parent
<kings_> adambeynon_: ah fixed it !
<kings_> it was paths
<meh`> adambeynon_, again, to be clear, the only difference now is Numeric has been renamed to Number, and Number needs Numeric as superclass
<meh`> that's all
<meh`> there are no other changes to how Number works
<kings_> adambeynon_: actually I take that back - it's still hanging - but it's trying to run them at least now - must be to do with phantomjs
<meh`> this allows us to implement Decimal, Complex, Rational, BigNumber and anything number related
<kings_> well I got the example to work so getting there slowly
<kings_> well I can only getting working if I put the specs in a subdirectory
<kings_> also can I read from file?
<kings_> in my test?
<kings_> I want to pass some json to my class under test
<kings_> in the main app - I get the json from an HTTP.get
<kings_> but I wanted to put the test json in a file and grab it from there
<meh`> kings_, nope
<kings_> do I have to use HTTP.get in my test file then
<meh`> kings_, you can setup the spec server to deploy the file
<meh`> yes
<kings_> so I have to require 'opal-jquery' in my test file as well as the jquery lib
<meh`> kings_, yeah, but, why does it have to be a file?
<kings_> its large
<meh`> well
<meh`> you could use the __END__ block
<meh`> how large is it?
<kings_> 7000 lines
<meh`> mh
<meh`> maybe there's a way to compile it in the specs
<meh`> adambeynon_, ^
<kings_> its ok I got it working from a file
<meh`> adambeynon_, anyway, it's not enough to hack the superclass in number
<meh`> we need to fix bridged classes
<meh`> gh476 is caused by them bein broken
<[o__o]> Wrong method calling on extended bridged classes:
kings_ has quit [Quit: Leaving...]
<adambeynon_> kings_: you could stub HTTP.get()
<adambeynon_> 5 seconds too late :|
<adambeynon_> kings_: if your back later..
<adambeynon_> meh`: is securerandom suitable for for guid's ?
kings has joined #opal
<meh`> yes
GitHub58 has joined #opal
GitHub58 has left #opal [#opal]
<GitHub58> opal-haml/master 5e6453c Adam Beynon: Specs can be run in browser
<GitHub58> [opal-haml] adambeynon pushed 1 new commit to master:
<kings> if I have json as a string - how can I convert it into a ruby object that is a json object?
<adambeynon_> kings: JSON.parse(string)
<kings> parsing fails
<meh`> kings, then syntax errors in the JSON
<adambeynon_> kings: the json parse uses the standard browser JSON object
<meh`> JSON.parse uses the browser parser
<kings> if I replace ":" with " => " - then I can use the object without it being a string
<kings> but I cant seem to parse it when its a string
<adambeynon_> kings: it is json yes? not ruby as a string?
<kings> yeah it json
<kings> but if I convert it to a ruby object by doing that replace from : to => then it works as an object
<kings> but as soon as I put it in json=<<-EOF json EOF - then it wont parse
<kings> theres is not much stack trace info available - just says failed to parse string
<kings> I've found it pretty hard to get my opal-rspecs working as there is hardly any error info coming out
<kings> is there a way to get a full stack trace?
<adambeynon_> kings: it should show the backtraces for each failed spec
<adambeynon_> kings: which browser?
<kings> its using phantomjs - running as a rake task
<kings> bundle exec rake
<kings> its a test
<adambeynon_> kings: it should still be showing a stack trace for each failed spec
<adambeynon_> kings: can you copy-paste the exact error message?
<kings> 1) ReportParser should parse a cucumber json file and return features
<kings> Exception:
<kings> Unable to parse JSON string
<kings> thats it
<adambeynon_> kings: ok, thats definitely the native json parser from the v8 engine, so it must be malformed json somewhere
<kings> weird how it works as an object tho and not as a parsable string
<kings> it also works fine via HTTP.get(json_file)
<meh`> kings, certain exceptions aren't catchable
<meh`> and have no stacktrace
<meh`> browser limitation
<kings> hmm ok
<meh`> kings, but make the json go through a linter
<meh`> or dependin on how you embedded the json
<meh`> might be doing escaping that it shouldn't do
GitHub23 has joined #opal
<GitHub23> 85ff5e5 Adam Beynon: Initial rspec docs
GitHub23 has left #opal [#opal]
<GitHub23> [] adambeynon pushed 1 new commit to master:
<kings> ok - I'm trying to find the root of the issue - it's pretty tough - probably a combination of my lack of familiarity with opal and not having much info in the stack
DrShoggoth has joined #opal
<kings> oh its much easier running the specs in the browser with debug=true
DrShoggoth has quit [Client Quit]
DrShoggoth has joined #opal
<meh`> kings, yeah, phantomjs is kind of a huge hack
<kings> ok figured it out at last - there was some newline characters amongst the json - JSON.parse didnt like it - when I replaced all "\n" with "" it started working
<kings> hell yeah! I'm dancing around the room with joy!!!
DrShoggoth has quit [Quit: Leaving]
<meh`> adambeynon_, I was thinking, instead of creating the classes in bridge_class, couldn't we bridge the native to an already existing class?
<meh`> and doing the bridging in the class itself?
DrShoggoth has joined #opal
<meh`> I mean, in the file instead of the runtime
<meh`> this would fix #476 and allow for Number < Numeric
<adambeynon_> meh`: we can already do something like:
<adambeynon_> class Array < `Array`; ... end
<meh`> adambeynon_, yeah but that wouldn't solve the issue
<adambeynon_> meh`: a bridged class cant be a subclass
<meh`> adambeynon_, why?
<adambeynon_> bridges only work when they are direct children of Object
<meh`> why?
<adambeynon_> otherwise it becomes really complex in donating methods between multiple source classes and target classes
<meh`> I see
<meh`> I guess #476 comes from broken donation then
<meh`> it just overrides blindly
<adambeynon_> could be - haven't looked at it in detail yet
<meh`> adambeynon_, +1 on that post
<meh`> most people don't understand the difference between syntax and semantics
<meh`> they think everything is about syntax
<adambeynon_> meh`: yep. I think we could do with a nice intro part on these new docs, sort of design deciscions
kings has quit [Quit: Leaving...]
<adambeynon_> "why cant I call my fav_lib.js" etc
<adambeynon_> I think it will also filter users straight away, who opal is for, and who it isnt
<adambeynon_> adambeynon_: I was half thinking of writing a small angularjs style clone, to deal with directives in the html
kings has joined #opal
<adambeynon_> but after watchign a screencast, I never wish to do anything of the sort
<adambeynon_> it is awful
<adambeynon_> I dont get how people could write their gui apps like that
<meh`> yeah
<adambeynon_> half of it is inline javascript within the html
<meh`> yep
<adambeynon_> didnt we decide like 10 years ago that ujs was a good idea?
<meh`> we could go "want backbone? use vienna; want angular? use lissio"
<meh`> or something of the sort
<adambeynon_> meh`: yep. I think a big (bigger) demo app is really needed too
<meh`> adambeynon_, working on that, gwentoo
DrShoggoth has quit [Read error: Connection reset by peer]
<adambeynon_> meh`: the first angular demo I found tells me that this code is "a clean solution":
<adambeynon_> <body><a ng-click="model.set({selected: true})"></body>
<meh`> lol
<meh`> isn't angular made by google?
<meh`> everything they do is awful
DrShoggoth has joined #opal
<meh`> Dart, Go, libraries, they're all absolutely disgusting
<meh`> adambeynon_, anyway, right now my biggest issues with Opal now are Numeric and require semantics
<meh`> without them I can't run the gems I'm interested in
<adambeynon_> meh`: right now I am digging through the sprockets code base to see if we can get an implicit opal.main()
<meh`> and they're VERY common things in gems
<adambeynon_> meh`: a REALLY ugly way would be to keep a list of requires inside our processor
<adambeynon_> i.e., if an asset was required somewhere, wrap it in a module thing
<adambeynon_> otherwise assume it was top
<adambeynon_> or, if an asset was required using a ruby require, rather than a sprockets comment, then wrap it
<meh`> sincerely I don't care how ugly it is as long as it works, it's too much of an important feature
kings has quit [Quit: Leaving...]
kings has joined #opal
kings has quit [Ping timeout: 272 seconds]
elia has joined #opal
GitHub120 has joined #opal
<GitHub120> ab6fe73 Adam Beynon: Some docs reorganization
<GitHub120> [] adambeynon pushed 1 new commit to master:
GitHub120 has left #opal [#opal]
kings has joined #opal
kings has quit [Ping timeout: 248 seconds]
<meh`> adambeynon_, shouldn't __hash and __range be $hash and $range?
elia has quit [Quit: Computer has gone to sleep.]
marcandre has joined #opal
<meh`> adambeynon_, I don't see the issue with bridged classes having a non-Object superclass
<meh`> it's just a matter of cleaning up defn and donate to be recursive instead of hardcoded
<meh`> adambeynon_, or am I missing something?
kings has joined #opal
<meh`> adambeynon_, I think we need a an Opal.def helper
<meh`> so we can fix #476 and allow for more fine grained donation, and optionally support method_added as well
<adambeynon_> meh`: we do
<adambeynon_> Opal.defn()
<adambeynon_> and Opal.defs()
<adambeynon_> method_missing on bridged classes?
<meh`> no no
<meh`> non-Object superclass
<meh`> and #476
<meh`> overriding the methods
<meh`> isn't defn for module methods?
<meh`> or something
<meh`> if I look at the produced opal.js
<meh`> most methods are defined with just def.$method = whatever
<meh`> instead of going through a helper like the rest
<adambeynon_> meh`: it is generally used for modules and ::Object and ::BasicObject
<adambeynon_> it *can* be used with any other class with no side-effects
<meh`> is there a reason we don't use it everywhere?
<adambeynon_> readability of generated code I guess
<adambeynon_> and less overhead
<meh`> looks less readable to me sincerely
<meh`> def.$chars = TMP_1 = function() {
<meh`> def['$=~'] = function(other) {
<meh`> and I don't think the overhead would be that noticeable, that kind of call pattern is optimized out
<meh`> and we will need it anyway to support method_added
<meh`> mmh
<adambeynon_> vs: Opal.defn(def, '$=~', function(other) ...
<meh`> yeah, why does defn and defs take the prefixed method name?
<meh`> it should take the mid instead of jsid
<meh`> meh
<meh`> I'll work on a separate branch
<meh`> and try to fix #476
<meh`> it's time I dive into the runtime
<meh`> two people who can work on it are better than one
DouweM has quit [Ping timeout: 245 seconds]
<meh`> adambeynon_, while I'm there I'm starting the cleanup as well, I think we should call special variables $$name or $name$, _ is usable in ivar names
GitHub167 has joined #opal
<GitHub167> opal/master 8514670 Adam Beynon: Opal::Server should not search for an index.html path unless specified
<GitHub167> [opal] adambeynon pushed 1 new commit to master:
GitHub167 has left #opal [#opal]
kings has quit [Ping timeout: 245 seconds]
kings has joined #opal
travis-ci has joined #opal
<travis-ci> [travis-ci] opal/opal#1680 (master - 8514670 : Adam Beynon): The build passed.
<travis-ci> [travis-ci] Build details :
travis-ci has left #opal [#opal]
GitHub101 has joined #opal
<GitHub101> [opal-todos] adambeynon pushed 1 new commit to master:
<GitHub101> opal-todos/master 57733de Adam Beynon: Add some specs for example
GitHub101 has left #opal [#opal]
ryanstout has joined #opal
<meh`> adambeynon_, what's the difference betwee _super and __parent?
<adambeynon_> __parent includes modules and singleton classes
<adambeynon_> _super is the actual .superclass()
<adambeynon_> meh`: dont swap to $$foo private vars
<meh`> I see
<adambeynon_> I dont like the double $$
<meh`> yeah it looks ugly
<meh`> already tried
<meh`> trying to think of a good way
<meh`> a .$ special table doesn't look half bad, but I'm thinking about the repercussions
<meh`> keeping in mind that if we optimize the shape, it's going to be blazingly fast
<meh`> so ignore the idea of overhead there
<adambeynon_> special table for private variables?
kings has quit [Quit: Leaving...]
<meh`> if by private variables you mean the various special flags, yes
<adambeynon_> it would make it a pain to debug opal problems/speed/bugs
<meh`> adambeynon_, how so?
<adambeynon_> well, instead of inspecting an object in the debugger, I would have to keep opening the variable table as a second object
<adambeynon_> obj._klass vs obj.table._klass
<meh`> well, obj.$.klass
<meh`> any other ideas?
<meh`> the current situation is unbearable
<adambeynon_> meh`: is it currently that bad?
<meh`> adambeynon_, yes
<adambeynon_> rubinius uses ivars for some runtime stuff too
<meh`> _a, __b, __c__
<adambeynon_> meh`: well, I can make them more consistent
<adambeynon_> __foo
<meh`> adambeynon_, but then we get into a vicious cycle
<meh`> you went with __parent because rspec was using @_parent
<meh`> we should use a character unusable in ivars
<meh`> also I think Rubinius usage of ivars for runtime stuff ends up creating weird bugs
<meh`> when inheriting from core classes mostly
dragonnkh has joined #opal
dragonkh has quit [Read error: Connection reset by peer]
marcandre has quit [Remote host closed the connection]
kings has joined #opal
<kings> evening
<meh`> evening
<kings> hey meh`
<kings> meh`: my opal project is a single page app - but I'm not sure exactly how to display it - so I consume some json which then gets turned into an object hierarchy - I then have 3 views I'd like to show - first an overview - the json is made up of features with scenarios and steps - so the overview shows an overview of all the features and their pass/fail status and some graphs
<kings> the 2nd view - is of each specific feature - listing the scenarios and steps for the feature
<kings> the 3rd view is showing all the features but aggregated by tag (Each feature has a tag)
<kings> so my question is - should I have 3 links on the index page - which when clicked - replace the view with the appropriate view? and when a link is clicked - should I at that point generate the html and render it?
<kings> or should I generate and render everything - and have the links just show and hide various views?
<kings> any advice welcome
<meh`> are you using any framework?
<kings> no I wanted to do it without as a learning experience
<kings> just opal-jquery
elia has joined #opal
<kings> hey elia
<kings> kings/dragonkh here :)
<meh`> kings, I'd render on click
<elia> hey there :)
<kings> yeah I think thats a good idea
<elia> k, checking transcript :)
<kings> :)
<kings> elia: I've had some real fun trying to get my opal-rspec working today ! - but its working now :)
<elia> kings, should write some docs down, while there's still pain :)
<kings> hehe yeah - I will - although some of it was more related to my dense brain than anything else hahah
<elia> meh`, adambeynon_, ng-lol
<meh`> lol
<adambeynon_> kings: I found a slight bug with opal-rspec earlier where it was trying to load any index.html file it could fine
<adambeynon_> Find in your app
<adambeynon_> It caused my specs to hang
<adambeynon_> So that might have been the same issue, maybe
DouweM has joined #opal
<kings> perhaps - I had to move my spec folder into a subfolder for it to work - otherwise it just hangs - there is an index.html at the root - level - so it could be the issue
<adambeynon_> kings: I think so.. I have fixed it on opal master branch which has resolved the issue for me
<kings> ok great :) I'll try it later
<elia> adambeynon_, got anything out of sprockets?
<elia> adambeynon_, I was going to have a look too, but maybe I'm late
kings has quit [Quit: Leaving...]
<adambeynon_> elia: I did, but didn't get too far. It looks like a top level asset and a required asset take the exact same route in sprockets.
<adambeynon_> We would have to introduce any solution ourselves
<adambeynon_> Which is tricky
<elia> adambeynon_, was thinking the same thing
<elia> adambeynon_, probably the way is to avoind passing parsed requires to sprockets and just prepend them while compiling
<elia> `compiler.requires.each do |r|`…
<elia> and now that im looking at it it already looks for rb files
<elia> in #find_opal_require
<elia> the comment requires can still go to sprockets safely
<elia> if we find a way to do it cache-wise I think we're done
mneorr has joined #opal
adambeynon_ is now known as adambeynon
<elia> adambeynon, YAY, _ no more!
<adambeynon> elia: trying to msg NickServ to change it with an iPad keyboard is really difficult :p
<adambeynon> elia: hijacking the requires could do the trick I think
<elia> adambeynon, the other alternative I was thinking of is some hack with filenames or mime types
<elia> inlining requires is what sass does when you use @import
<elia> adambeynon, the downside is that it won't use separate script tags (in rails at least)
<elia> adambeynon, sass_importer.rb seems interesting too
<adambeynon> elia: right, havent noticed the @_dependency_paths variable before
<adambeynon> elia: if there was some @_dependant_assets, then we would be ready to go
<elia> adambeynon, me too
<elia> seems that sass had the same kind of problem
mneorr has left #opal [#opal]
mneorr has joined #opal
elia has quit [Quit: Computer has gone to sleep.]
elia has joined #opal
mneorr has quit [Remote host closed the connection]
mneorr has joined #opal
mneorr has quit [Ping timeout: 272 seconds]
kings has joined #opal
kings has quit [Client Quit]
mneorr has joined #opal
marcandre has joined #opal